A novel threshold-based scan detection method using genetic algorithm

Abstract

In order to attack to a network, an attacker first must find vulnerability points of the target network. This task is done through scanning. There are many methods of scan detection. Most of these methods are based on thresholding. Setting a proper threshold value is crucial and depends on many parameters such as network structure and time window. In this study we proposed a new scan detection method based on genetic algorithm (GA). This method has two phases. In the first phase we separate normal traffic from suspicious traffic and send only suspicious traffic to the second phase. This way the overhead of the process in the second phase is decreased considerably. In the second phase we aim to detect attacks with respect to two optimum parameters of threshold and memory. We compared our method with snort. Results showed that our method achieves better performance in both hit rate and false alarm rate.